Refining the Shortest Paths (RSP)

2.1. Preliminary analysis using actel

Analysing acoustic telemetry data requires initial filtering to exclude misleading data (e.g. false detections, detections prior to release). To overcome this issue and ensure reliable results, the RSP toolkit operates in close relationship with the actel R package, which is filters and invalidates flawed detections. Before getting started with RSP, you will have to download actel and filter your acoustic data. Please click here for more information and to download actel.

To start using RSP, you can run the simplest actel analysis with the function explore. You can find more about how organize your data and run this preliminary analysis in actel’s manual pages (run ?actel after loading the library).

It is important that you save the output of the actel function, so you can later on use it to calculate the RSP. e.g.:

library(actel)
filtered_data <- explore(tz = "Europe/Sydney")

2.2. Creating and exporting a raster file from the study area in QGIS

After filtering your acoustic data, you will need a raster file from your study area defining the water and land limits. This file will be used for estimating the shortest paths in water between consecutive acoustic detections, using a least-cost analysis of constrained random walks. The values of the raster cells must comprise zeros (water) or ones (land). Depending on the size of your study area, a resolution of 0.0001 (latitude ° x longitude °) is suggested for more accurate estimations (especially for sites with very narrow channels). Please see the following steps for generating and exporting a raster file from you study site in the QGIS software:

1. Load a good-resolution shapefile of your study region into QGIS:

2. Zoom into your study site and create an overlapping polygon that encompasses the entire area,

3. Clip your study site from the shapefile using “Vector > Geoprocessing Tools > Clip” setting the shapefile as the input layer and the overlap polygon as the overlay layer:

4. Create a raster layer from the Clipped shapefile using the “Rasterize” tool from the toolbox setting the Output Values to “data / no-data” and Cellsize to 0.0001 for a better resolution:

5. Right-click on the raster layer and select “Export > Save As…”. In the new window select the “Golden Software Binary Grid (.grd)” from the dropdown format menu. In the extent menu, select the “Calculate from Layer” option and then click on the “Clipped” layer:

6. You can now import the raster from your study area into R and plot it to check whether the resolution (Cellsize) chosen is good enough:

library(raster)
LakeMacquarie <- raster("Lake_Macquarie.grd") # Import the raster file exported from QGIS
plot(LakeMacquarie) # Plot raster from the study area

If the river channels look distorted go back to step 4 and choose a lower Cellsize. Please be aware* that increasing the raster resolution too much will require higher computational costs and may cause R to crash.

Now that your acoustic detections have been filtered and that you have a raster file with good resolution from your study area (exported into your working directory), you are all set to get started with RSP.

3.1. Exploring the RSP results

Here are some examples of the runRSP() output:

1. For consecutive detections on the same receiver:

The Position column in this dataset identifies the two consecutive acoustic detections (Receiver) from this animal. We can notice that they occurred on the same Receiver (125449): the first on 2018-03-07 00:43:49 and the second on 2018-03-07 01:11:45 (approximately 30 minutes from each other). Because this time difference is longer than the default time.lapse (10 minutes), the runRSP() estimated the intermediate positions (RSP) by repeating the receiver Longitude and Latitude and changing the Error parameter (at a rate of 5% from the default distance argument of 250 meters = 12.5 meters).

2. Consecutive detections on different receivers:

Here the animal was detected first at the Receiver 100474 on 2018-04-27 05:27:10, and then at the Receiver 107527 on 2018-04-27 07:04:39. The runRSP() now calculated the shortest in-water path between receivers, and we can see how the Error of added locations increased up to half-way, (575 meters on 2018-04-27 06:15:54), and then decreased back to 500 as the track approached the second receiver.

3.2. Visualizing RSP outputs

We can use plotDistances() to compare the total distances travelled by each animal calculated using only the receiver locations and or also including the RSP estimations:

The plotDetections() shows the total number of receiver and estimated positions for each tracked animal:

You can also plot all tracks from a particular animal using plotRSP():

# Plot tracks only with receiver locations:
plotRSP(input = rsp.results, tag = "R64K-4075", display = "Receiver", type = "lines") 

# Plot tracks with RSP estimations:
plotRSP(input = rsp.results, tag = "R64K-4075", display = "RSP", type = "lines")

You can also set display = "Both" to plot both track options on a single plot.

4.1. Total dynamic Brownian Bridge Movement Model (group dBBMM)

This option calculates a series of dBBMM for each animal track from all the groups monitored. The breaks argument defines for which contours the areas of use should be calculated, which by default are the 50% and 95% (i.e. breaks = c(.5, .95)).

Track quality checks are performed to ensure that only good tracks which allow the dBBMM to converge are included in the analysis. This is an example of the returned messages from dynBBMM():

> group.dbbmm.results <- dynBBMM(input = rsp.results, UTM.zone = 56, breaks = c(0.5, 0.95))

M: Preparing data to apply dBBMM.
M: No specific transmitters selected. All the data will be used for analysis.
Warning: 8 track(s) in group Bream have less than eight detections and will not be used.
Warning: 5 track(s) in group Luderick have less than eight detections and will not be used.
Warning: 1 individual detections were removed in group Tarwhine due to simultaneous detections at two receivers.
Warning: 2 track(s) in group Tarwhine have less than eight detections and will not be used.
M: In total, 57 detections were excluded as they failed the track quality checks.

After calculating UDs, the land areas are excluded so that the final results represent only in-water areas of use. The overall overlap between each group monitored is also calculated.

M: Subtracting land areas from output.
M: Calculating overlaps between groups.
M: Storing final results.

The results of the dBBMM are saved in the $track.areas object, as a list of data frames for each group analysed:

Each Track is named after the transmitter and the corresponding track name and both Start and Stop timestamps are stored. Areas of use (by default Area.5 and Area.95) are saved in squared meters, together with the respective elapsed times in minutes (Time.lapse.min). Note that in this example for the animal R64K-4075 only the Track_3, Track_7 and Track_8 were used for calculating space use areas, as the remainder ones failed the quality checks.

You can use plotContours() to visualize any of the dBBMM calculated by specifying the group and track you want to plot:

plotContours(input = group.dbbmm.results, group = "R64K-4138", track = 'R64K-4138_Track_7', main = "R64K-4138_Track_7")

The levels argument can be used to specify which contours to plot.

plotContours(input = group.dbbmm.results, group = "Bream", track = 'A69-9002-10473_Track_08', main = "A69-9002-10473_Track_08", stations = TRUE)

If stations = TRUE the locations of receiver stations are added to the plot.

4.2. Total space-use overlap

(REVIEW STOPPED HERE!)

Now that we calculated the areas of space-use within our study area for each group monitored, we can investigate the ammount of overall overlap between the different species stored in the $overlap.areas object:

group.dbbmm.results$overlap.areas$`0.5`
group.dbbmm.results$overlap.areas$`0.5`$absolute

group.dbbmm.results$overlap.areas$`0.5`$percentage

group.dbbmm.results$overlap.areas$`0.95`
group.dbbmm.results$overlap.areas$`0.95`$absolute

group.dbbmm.results$overlap.areas$`0.95`$percentage

Please note the overlaps are calculated for the contours defined by breaks in dynBBMM(), and returned both in absolute values (squared meters) and percentage matrices. For example, we can see in the last table that Bream and Luderick were the groups with higher overall overlap of 63.36% at the 95% level, whereas Tarwhine and Luderick only had a 38.41% overlap at the 95% contour. To see exactly where space use overlaps occurred you can use plotOverlap():

plotOverlap(input = group.dbbmm.results, stations = FALSE, level = .95, store = TRUE)

4.3. Fine-scale dynamic Brownian Bridge Movement Model (timeslot dBBMM)

dBBMMs can also be calculated according to a moving temporal window. This allows investigating how the space-use overlap between the different groups varied during the study period. It is useful for assessing the influence of environmental parameters upon space-use of different groups tracked within the study area. The same dynBBMM() function is used, but here the argument timeframe has to be defined in hours as the temporal window. The total tracking period will be divided into timeslots, and dBBMMs calculated for each group monitored (for each timeslot). Overlapping areas are now calculated for each timeslot and the corresponding metadata stored in the $timeslots object:

time.dbbmm.results <- dynBBMM(input = rsp.results, UTM.zone = 56, breaks = c(0.5, 0.95), timeframe = 12) # 12-h timeslots
head(time.dbbmm.results$timeslots)

In the example above we can notice that in the first timeslot none of the groups were detected, whereas Luderick group got detected in slot 2 and both Luderick and Tarwhine were detected on the first 12-h of 2013-09-03 (slot 3). The $track.areas object for each tracked group will now have a first column named Slot, which identifies the timeslot for each of the dBBMM calculated:

head(time.dbbmm.results$track.areas)?

what was the command to get the table below exactly?

Here we can see that two animals were detected for this particular group between Slot 486 and Slot 488: the transmitters A69-9002-10474 and A69-9002-10480. The Time.lapse.min will now be very similar as its maximum elapsed time is limited by the timeframe argument, and it depends on the times of first and last locations for each animal.

The following command line can help you assess if any other group got detected during this same timeslot:

> time.dbbmm.results$timeslots[486, ]
    slot               start                stop Bream Luderick Tarwhine
486  486 2014-05-02 12:00:00 2014-05-03 00:00:00  TRUE     TRUE    FALSE

Yes, both Bream and Luderick were detected between 2014-05-02 12:00:00 and 2014-05-03 00:00:00. We can inspect whether the two groups overlapped or not:

> time.dbbmm.results$overlap.areas$`0.95`$percentage$`486`
             Bream  Luderick Tarwhine
Bream           NA 0.8247734       NA
Luderick 0.8247734        NA       NA
Tarwhine        NA        NA       NA

This shows that the two groups had an overlap of 82.47% at the 95% dBBMM contour during this particular timeslot. We can now see exactly where the overlap occurred by plotting the space-use models using:

# plotContours() for plotting the individual dBBMM:

plotContours(input = time.dbbmm.results, track = "A69-9002-10474_Track_1", group = "Bream", timeslot = 486, stations = TRUE, main = "A69-9002-10474 (Bream - slot 486)")

plotContours(input = time.dbbmm.results, track = "A69-9002-10480_Track_1", group = "Bream", timeslot = 486, stations = TRUE, main = "A69-9002-10480 (Bream - slot 486)")

plotContours(input = time.dbbmm.results, track = "A69-9002-10481_Track_1", group = "Luderick", timeslot = 486, stations = TRUE, main = "A69-9002-10481 (Luderick - slot 486)")

# plotOverlap() to investigate the overlap areas:

plotOverlap(input = time.dbbmm.results, store = TRUE, stations = FALSE, timeslot = 486) 

Approximately 1-month of fine-scale (12-h) space use of Bream and Luderick at the 95% level